Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 1

LAT Light Curve Analysis

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 2

Photometry

LAT Light Curves can be obtained in two basic ways:– Likelihood Analysis– Aperture Photometry

Likelihood analysis has the potential for greatersenstivity. However, aperture photometry is easier,faster, and has the benefit of model independence.

This presentation only deals with aperture photometry

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 3

Aperture Photometry Process

It is recommended to use a script to create an analysischain from the toolset– fkeypar - extracts start and stop times– gtselect - performs desired data cuts– gtmktime - creates good time intervals– gtbin - makes quasi-light curve (counts rather than rate)– fdump - exports data– other tools - convert counts to rates, calculate errors

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 4

Extract Start/Stop Times

$ fkeypar “L090923112502E0D2F37E71_PH00.fits[1]” TSTART(photon start time = 266976000.)

$ fkeypar “L090923112502E0D2F37E71_PH00.fits[1]” TSTOP(photon stop time = 275369897.)

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 5

Perform Data Selection$ gtselect zmax=105 emin=100 emax=200000L090923112502E0D2F37E71_PH00.fits outfile=temp2_1DAY_3C454.3.fitsra=343.490616 dec=16.148211 rad=1 tmin=266976000. tmax=275369897.evclsmin=3 evclsmax=10

Parameters Specify– Energy Range (100 to 200,000 MeV)– Source coordinates– 1 degree aperture radius– Previously determined start and stop times

(Note: if you’re going to barycenter then the min and max times should beslightly greater/less than the times in the spacecraft file

– Event class min/max (3 for diffuse class, 0 for simulated data)

Writes to file: temp2_1DAY_3C454.3.fits

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 6

Calculate GTIs$ gtmktime scfile="L090923112502E0D2F37E71_SC00.fits"filter="(DATA_QUAL==1) &&(angsep(RA_ZENITH,DEC_ZENITH,343.490616,16.148211)+1<105) &&(angsep(343.490616,16.148211,RA_SCZ,DEC_SCZ)<180)" roicut=nevfile=“temp2_1DAY_3C454.3" outfile="temp3_1DAY_3C454.3"

Parameters Specify– Good data quality– Photons less than 105 degrees from zenith (+1 is for 1 degree aperture)– Photon locations not in the exact center of the field of view– Input file is output from gtselect

Writes to file: temp3_1DAY_3C454.3.fits

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 7

Extract a Light Curve$ gtbin algorithm=LC evfile=temp3_1DAY_3C454.3.fitsoutfile=lc_1DAY_3C454.3.fitsscfile=L090923112502E0D2F37E71_SC00.fits tbinalg=LINtstart=266976000. tstop=275369897. dtime=86400

Parameters Specify– Make a light curve (LC)– Input file is output file from gtselect– Spacecraft file– Linear time bins– Start and stop times– Size of time bin in seconds (86400 = 1 day bins)

Writes to file: lc_1DAY_3C454.3.fits

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 8

Calculate Exposures$ gtexposure infile="lc_1DAY_3C454.3.fits"scfile="L090923112502E0D2F37E71_SC00.fits" irfs="P6_V3_DIFFUSE"srcmdl="none" specin=-2.1

Parameters Specify– Spacecraft file– Instrument Response Function to use– Source model (for more complex model than power-law)– Spectral Index for use in power-law spectrum (convention requires minus sign)

EXPOSURE column is added to file: lc_1DAY_3C454.3.fits

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 9

ResultsFinal FITS file contains:– Time in MET– Bin Width in seconds– Number of counts per bin– Error– Exposure

To convert to rates, use fv or other tool to divide countsby exposureError bars in output are sqrt(counts)

In some instances (e.g. too few counts) this may be incorrectCorrecting this may be more complicated

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 10

Light Curve

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 11

Barycentering• If your source is sensitive to the motion of the Earth, you may wish to

barycenter the events file to remove that effect

• gtbary is usually used to barycenter the events file for pulsar timing. But itcan also be used for light curves

– gtbary must be the last step of the analysis (after exposure caluclation)

– Spacecraft file must be longer than the events file (remember this when doingthe gtselect step)

– gtbary overwrites the time column with the barycentered (corrected) photonarrival times. It’s wise to make a copy of your data file before running gtbary.

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 12

Using Exposure Errors• For some purposes, errors based on observed counts may not be correct

• Alternative is to use errors based on the exposure– Calculate the mean count rate

– For each time bin, calculate the expected number of counts based on theexposure for that time bin

– Take the square root of that predicted number of counts

– Divide by the exposure to get the rate

– The resulting error value is based only on the “quality” of each time bin

• References for error bars treatment:– Gehrels, 1986, ApJ, 303, 336

– Kraft, Burrows, & Nousek, 1991, ApJ, 374, 344

Fermi Analysis Workshop Boston University, Jan. 11, 2010 Elizabeth Ferrara - FSSC 13

Comparing Error types